J Geograph Syst (2004) 6:155–174

DOI: 10.1007/s10109-004-0131-6

Integrating Web-based GIS and image processing tools

for environmental monitoring and natural resource
management
Ming-Hsiang Tsou
Department of Geography, San Diego State University, San Diego, CA 92182-4493, USA
(e-mail: mtsou@mail.sdsu.edu)

Abstract. The combined powers of Web-based geographic information

systems (GIS) and on-line remote sensing tools can significantly reduce the
high cost and labor associated with environmental monitoring and natural
resource management. This paper introduces an integrated Web-based GIS
architecture by combining three levels of geographic information services
(GIServices): data archive, information display, and spatial analysis. A
prototype Web site, WGAT (Web-based GIS and Analytic Tools), has been
developed to provide easy access of geospatial information and to facilitate
Web-based image analysis and change detection capabilities for natural
resource managers and regional park rangers. The Web-based integration
framework emphasizes user-oriented services, distributed network environ-
ments, metadata standards, communication protocols, client/server compu-
tation, and ubiquitous access.

Key words: Web-based GIS, remote sensing, GIServices, environmental

monitoring

1 Introduction

The advent of the Internet and Web-based geographic information systems

(GIS) technologies provides a convenient and efficient way to access and
disseminate geospatial data and remotely sensed imagery. There is a great

This paper forms a portion of the ‘‘Integrated Mobile GIS and Wireless Image Web Services for
Environmental Monitoring and Management’’ project supported by NASA’s Affiliated Research
Center (ARC) at San Diego State University. Funding by the NASA ARC program and
matching funds from the San Diego State University Foundation are acknowledged and greatly
appreciated. The author wishes to thank John Kaiser, the ARC program coordinator, and
Dr. Douglas Stow, the ARC program Principle Investigator, for their coordination efforts on
this project. Java programming was provided by Liang Guo, a geography graduate student at
San Diego State University.
156 Ming-Hsiang Tsou

potential for using Web-based GIS and on-line image analysis tools in the
areas of natural habitat preservation and environmental monitoring. The
combined powers of data collection through remote sensing techniques and
geospatial analysis tools by means of the Internet can significantly reduce the
high cost and labor associated with traditional field monitoring and
environmental resource management methods.
However, current development of Web-based GIS mainly focuses on
vector-based geospatial data rather than raster-based remotely sensed
imagery. This paper argues the importance of integrating raster-based image
data and vector-based feature data under a single Web-based framework
from a software developer’s perspective. An integrated Web-based GIS
architecture was introduced by combining three levels of Web-based
geographic information services (GIServices): data archive and search,
information display and query, and spatial analysis functionality. A
prototype system, WGAT (Web-based GIS and Analytic Tools, http://
map.sdsu.edu/arc) has been developed to provide easy access of geospatial
information and to facilitate the adoption of image analysis and change
detection methods for natural habitat management and monitoring. A three-
level Web service framework was created to monitor multi-species habitat
and land cover changes for the Mission Trail Regional Park (MTRP) located
in San Diego, California. Figure 1 illustrates the three levels of services and
the actual implementation of the prototype Web site.
The first level of integration is the data archive/search service. A Web-
based data warehouse was developed for archiving, accessing and down-
loading both GIS databases and remotely sensed imagery. The major
integration requirement is the adoption of a standardized metadata format
for both GIS layers and remotely sensed imagery. The Web-based data
warehouse can help users to index and search the contents of metadata.
The second level is the information display and query service. Multiple
interactive map servers and image servers were established to provide Web-
based mapping functions for the display of land use, vegetation, soil, trails,
roads, and remotely sensed satellite images. The major integration require-

Level of integration Implementation and requirements

Level three: On-line spatial analytical tool

Spatial analysis
Ubiquitous access

Client/server computation balance.
æ
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Level two: Internet map/Image server

Information display and query
Web-based display

Communication protocol
æ
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

Level one: Web-based data warehouse

Data archive and search
Standardized metadata

Metadata index/search

Fig. 1. The three-level integration of Web-based GIS and on-line remote sensing tools
Integrating Web-based GIS and image processing tools 157

ment at this level is providing an effective Web-based display mechanism and

a client/server communication protocol
The third level is the spatial analysis service. Several Java-based on-line
analytical tools (Java applets) were developed to provide advanced image
comparison capabilities and functions for land cover change detection
analysis. The implementation of Java applets can provide ubiquitous access
for natural resource managers and park rangers. The design of Java runtime
environments can create a balanced client/server computation framework.
There is an interdependent relationship between each service level. From a
system implementation perspective, the lower level services are the
prerequisite of higher level services. For example, to create a Web-based
map display and query (level two), the Web server has to connect to a
completed GIS database with standardized metadata (level one). The
functions of metadata index/search in level one will also be used in level
three (spatial analysis) for advanced GIS functions and geocomputations.
From a user’s perspective, the higher level services are the extended
transform of the lower level services. GIS users can download geospatial
datasets from data warehouse (level one), and then use their own GIS
software to perform map display (level two) or advanced spatial analysis
functions (level three) using their desktop GIS packages.
Under the Web-based framework, all three-levels of service were accessible
by using a standard Internet Web browser, which serves as the image viewer
and interface to a suite of image processing and GIS tools. The standard
Internet Web browsers and easy-to-use interfaces provide a flexible means to
access both spatial information and powerful geospatial analytical tools for
environmental monitoring tasks. The project research team worked closely
with the park rangers from the Mission Trail Regional Park in San Diego,
CA to develop and test the prototype analytical tools and demonstration
applications of this Web-based system. Park rangers and several GIS
professionals evaluated the prototype system, graciously provided their
expert review, and responded to the project team’s user-needs questions. The
project data sets include fifteen GIS layers from the San Diego Association of
Governments (SANDAG) and two types of remotely sensed imagery:
Airborne Data Acquisition and Registration (ADAR) system and IKONOS.
The ADAR 5500 system is an airborne multispectral digital camera system
with one meter spatial resolution. The IKONOS satellite is a high-resolution
commercial satellite system with one meter spatial resolution in panchro-
matic imagery mode and four meters spatial resolution in multipectral infra-
red imagery mode.
The key concept in the design of this prototype is to integrate Web-based
GIS and on-line remote sensing tools under a single framework. The
following section illustrates the challenges of combining these two types of
services and introduces related Web technologies and software.

2 Web-based GIS versus on-line remote sensing tools

The integration of GIS and remote sensing has been an important topic in
both GIS and remote sensing communities for a long time. Ehlers et al.
(1989) introduced a three level integration process between GIS and remote
sensing. The first level included simultaneous display of both vector and
158 Ming-Hsiang Tsou

raster data. The second level required dynamic data exchanges and display
functions via the same user interface. The third level of integration was
called total integration, which included both fundamental data model
integration and software development. Estes and Star (1993) introduced a
concept of integrated GIS (IGIS) that processed remotely sensed imagery
as well as raster and vector data sets in a consistent fashion. Fonseca et al.
(2002) added a fourth level of integration to Ehlers’s three-level framework.
The fourth level is the knowledge integration for GIS and remote sensing
applications. By using ontologies and a knowledge-based system, an
ontology-driven system can provide semantic integration of aerial images
and GIS and allow more flexible extraction of geospatial information
(Fonseca et al. 2002). These previous studies clearly indicate the essential
need to integrate GIS and remote sensing applications. Portions of this
paper and the proposed Web-based framework are inspired by these
studies. However, the three-level Web-based integration framework pro-
posed in this research is different from the system-based integration
mentioned in Ehlers’s article. The Web-based integration framework for
GIS and remote sensing tools developed in this project needed to consider
the nature of the distributed network environments, metadata, client/server
architecture, cross-platform programming tools, and software interopera-
bility. Also, the three-level integration framework emphasizes user-oriented
services rather than functionality-oriented systems.
The history of Web-based GIS can be traced back to the development of the
Xerox Map Viewer in 1994 (Putz 1994). The Xerox Map Viewer used a Web
Browser via HyperText Markup Language (HTML) format and Common
Gateway Interface (CGI) programs to provide interactive mapping functions
via the Internet. The technical framework of the Map Viewer was followed by
many early on-line GIServices applications. Another important research
project was the Alexandria Digital Library Project (Buttenfield and Good-
child 1996; Frew et al. 1998), which explored comprehensive services for on-
line spatial queries, map browsing, and metadata indexing. Distributed
component technologies and data/interface standards are also the foci of the
Internet mapping research, including OpenGIS specification (Buehler and
McKee 1998), ISO/TC211 Standards (Ostensen 1995), component-oriented
GIS (Li and Zhang 1997), and virtual data sets (Vckovski1998). Besides the
development of academic research, the GIS industry is also developing several
software packages to provide on-line mapping functions, such as ESRI’s
ArcIMS (Internet Map Server), INTERGRAPH’s GeoMedia Web Map,
AutoDesk’s MapGuide, and GE SmallWorld’s Internet Application Server.
In contrast, the early development of on-line remote sensing tools
mainly focused on data dissemination and catalog search functions. One
of the best examples is the Earth Observing System Data and Information
System (EOSDIS) developed by National Aeronautics and Space Admin-
istration (NASA) during the early 1990s. EOSDIS Version 0 (V0) system
was developed as an early prototype to provide an interoperable inventory
layer over existing, independent data systems. The main goal of EOSDIS
was to support the search and order access to heritage data collections
held by NASA’s Distributed Active Archive Centers (DAACs) (NASA,
2002). EOSDIS has been operating since August 1994 at eight Distributed
Active Archive Centers around the United States and interoperating with
Integrating Web-based GIS and image processing tools 159

six foreign sites. The early design of EOSDIS utilized the basic Internet
networking techniques, such as remote X-window display, UNIX operat-
ing systems, and TCP/IP communication mechanisms. After the advent of
the World Wide Web (the Web), the development of EOSDIS shifted to
Web-based technologies and the Earth Observing System (EOS) Data
Gateway (EDG) was created to provide multiple search functions for
remote sensing data available in EOSDIS. The current design of EOS
Data Gateway provides a Web-based access point for end-users to search,
query, and order remote sensing data. One unique feature of EOSDIS is
the introduction of an open architectural concept and the design of an
interoperability infrastructure between clients and data providers (Elking-
ton et al. 1994).
Different from Web-based GIS and Web-based mapping tools, most on-
line remote sensing applications mainly focus on cataloging functions and
graphic display. Very few applications provide advanced functions for
image analysis, such as the display of multiple spectral bands, georefer-
encing, image overlay, change detection, etc. Starting in the early 2000s,
along with the progress of Web technologies, including Java programming
and image compression techniques, the remote sensing industry began to
develop on-line remote sensing image servers and viewers, such as ER
Mapper’s Image Web Server (IWS) (with ECW viewer), Liztech’s Content
Server (with MrSID image format), and PCI’s Geomatica WebServer.
However, these software packages still lacked fully integrated capabilities
with other GIS Internet Mapping packages. The single exception is the
recent development of ER Mapper’s IWS plug-ins with ESRI ArcIMS
image server.
A major challenge to the full integration of the Internet map server with
the remote sensing image server is the heterogeneous architectures between
them. The different functionality required on the client side components
(viewers) is another major hurdle. Client/server problems are caused by the
three fundamental differences between Web-based GIS and on-line remote
sensing database sources: data archive formats, geographic information
display, and spatial analysis functions. To conquer these fundamental
problems, this paper suggests a fully integrated framework by combining
the three levels of services. The following sections illustrate the detailed
specifications of the three level integrations between Web-based GIS and on-
line remote sensing database sources. These services have been tested for
environmental monitoring and natural resource management tasks to
demonstrate their potential capabilities.

3 Level one: Data archive and search services

3.1 The development of data warehouse and metadata standards

The first level of integration between Web-based GIS and on-line remote
sensing database sources is the data archive and search services. The
Alexandria Digital Library and NASA EOSDIS Data Gateway, as
described earlier, are good examples of data archive and search services.
The Internet and the Web become the storage devices or media to archive
160 Ming-Hsiang Tsou

and deliver GIS data layers and remotely sensed imagery. The Web also
provides excellent user interfaces to catalog, index, and search these data
sets in the form of digital libraries. Two major forms of data archive and
search services are data warehouses and data clearinghouses. The role of
data warehouses (or data archive centers) is to archive data and to
provide data access, download, and preview mechanisms. Data clearing-
houses are built upon distributed metadata databases via multiple data
warehouses or other data clearinghouses (Tsou 2002). Current develop-
ment of data clearinghouses utilizes the Z39.50 protocol to index and
access multiple metadata repositories remotely. At present, Federal
Geographic Data Committee’s (FGDC) National Spatial Data Infrastruc-
ture (NSDI) and associated data clearinghouse nodes adopt this approach
to provide metadata search and query functions.
To create a successful data archive and search service, metadata is the key
issue for a full integration of both GIS layers and remotely sensed imagery.
Metadata can bridge the heterogeneous environments in GIS databases and
remote sensing data sets. However, the major challenge is to design a
comprehensive metadata standard for both GIS data layers and remote
sensing data. Currently, ISO 19115 Metadata Standard is the major
international metadata standard (previously published as ISO15046-15)
created by the International Organization for Standardisation (ISO) Tech-
nical Committee (TC) 211. The ISO metadata standards proposed a
conceptual framework and an implementation approach for geospatial
metadata that were developed partially based on the 1994 FGDC’s Content
Standard for Digital Geospatial Metadata (CSDGM) (FGDC 1998; ISO/TC
211/WG3, 1998). The major advantage of ISO 19115 and CSDGM
Metadata Standard is its flexibility in creating extensions and profiles for
various applications. With the ISO 19115 and CSDGM Metadata Standard,
the remote sensing community can define the metadata extensions for remote
sensing research and applications. In 2002, the extension for remote sensing
metadata was created and documented in the Content Standard for Digital
Geospatial Metadata: Extensions for Remote Sensing Metadata, FGDC-
STD-012-2002 (FGDC 2002). The following section introduces the actual
implementation of a data warehouse prototype combining both GIS layers
and remotely sensed imagery for environmental monitoring and manage-
ment tasks.

3.2 The data warehouse implementation

The project data warehouse was populated using selected GIS datasets from
the San Diego Association of Governments (SANDAG) archive and two
types (airborne and satellite) of remotely sensed imagery. Two series of
metadata are associated with the downloadable GIS layers and images
(Fig. 2). The first series of metadata was for remotely sensed images which
included 1998 (June 27, 1998), 1999 (May 26,1999), 2000 (May 3, 2000)
ADAR imagery, 1995 (June 23, 1995), 1997 (May 4, 1997) Digital
Orthophoto Quarter Quadrangle (DOQQ) imagery, and 2000 (June 29,
2000), 2001 (May 30, 2001) IKONOS multispectral imagery. The second
metadata series described the GIS themes which included 15 different map
layers, such as roads, trails, land use, soil types, vegetation, etc.
Integrating Web-based GIS and image processing tools 161

Fig. 2. The data warehouse implementation

Four major functions were provided by the data warehouse prototype:

metadata display, data preview, data download, and metadata search/ index.
The metadata of each GIS layer and remotely sensed image were displayed in
HTML format on the data warehouse pages. Each metadata file also
included a thumbnail image for preview of the actual data sets. The design of
the data warehouse also included a data download button to allow users to
download GIS layer exchange files and compressed remote sensed images
from the Web server. Another major function of the data warehouse was to
search the contents of metadata. The metadata search mechanism was
created to allow users to type keywords for searching and querying the
metadata. The search engine was implemented using Microsoft IIS Index
Services, a built-in function on a Window 2000 server. The search engine
indexed all metadata records stored in HTML format allowing efficient
metadata search and query functions.
162 Ming-Hsiang Tsou

In general, the implementation of data warehouse provides an easy-to-use

mechanism for resource managers and park rangers to access or download
GIS data and remotely sensed images that can be readily combined or
integrated with their own local GIS software projects. The data warehouse
prototype introduced here adopted an easy and direct approach for both
metadata contents and search mechanisms by utilizing existing software and
services (HTML documents and IIS index engines). On the other hand, there
are several potential limitations in the current design of data warehouse. For
example, the search mechanism only adopted text-based keyword matching
rather than semantic query and spatial-oriented operations. In the future, the
development of data warehouse, data clearinghouse, and metadata search
mechanisms will need to include more advanced methods for the full
integration of Web-based GIS and on-line remote sensing sources. The
adoption of new metadata extensions for remotely sensed data (FGDC
2002), eXtensible Markup Language (XML)-based information mediation
for metadata search and indexing tasks (Gupta et al. 1999), object-oriented
frameworks (Kemp 1999) and operational metadata schemes (Tsou 2002)
may establish more comprehensive and effective frameworks for data archive
and search services.

4 Level two: Information display and query services

4.1 Web-based display and communication protocols

The second level of integration between Web-based GIS and on-line remote
sensing sources is the information display and query services. Many
commercial software packages offer information display services by installing
their specialized map servers. However, the major challenges for the
integration of Web-based GIS and remote sensing tools are the compatibility
of data transmission/display format and client/server communication proto-
cols.
Traditionally, the display of GIS data mainly adopts the vector-based
data model as opposed to the raster-based model for remotely sensed
imagery. The same problem happens on the Web-based environments,
especially in the design of client-side viewers and server-side map engines.
The early development of Web-based GIS adopted the raster-based images
for map display. For example, the Xerox map server converted vector-
based layers into a GIF image (raster-based) on the server side and then
sent the GIF image to the client side in HTML format. The users viewed
the actual map as hyperlinked pictures on their Web browsers. More recent
development of Web-based GIS services began to adopt the vector-based
display format, such as Vector Markup Language (VML), Scalable Vector
Graphics (SVG), and Geography Markup Language (GML) (Zaslavsky
2000; OGC 2003). Vector-based display on the client-side viewers can
provide more comprehensive display functions than raster-based pictures,
such as rapid zoom-in/out, customizable map symbols, and layer stacking
order. The drawback of vector-based display is the incompatibility with
raster-based remote sensing data. Therefore, the major challenge for
information display services is to utilize both vector-based and raster-based
display on a single Web browser or viewer. One possible solution is to
Integrating Web-based GIS and image processing tools 163

redesign the client-side viewer so that it can access both vector-based map
servers and raster-based image servers at the same time. However, such a
function may require a ‘‘thick-client’’ solution, where the size of viewer
software modules will be larger than regular Web plug-ins or applets. It
will be difficult for users to download such large software modules via
current available network bandwidths.
Another challenge for information display and query services is the
design of the communication protocol between Web clients and servers.
To provide an interactive, dynamic map display, client-side viewers need
to send user’s request (such as zoom-in, zoom-out, add, query) back to
the servers. The servers will then generate new maps or information and
send them back to the client-side viewers. The design of client/server
communication protocol must make sure all the client requests can be
understood by the servers and vise versa. Currently, most GIS vendors
have their own proprietary communication protocols for Internet map
servers and image servers. For example, ESRI’s ArcIMS utilizes ArcXML
for its client/server communication protocols. ER Mapper’s Image Web
Server uses ECW Protocol (ECWP) for accessing ECW compressed
images and Web image servers. These vendor-based, specialized protocols
may cause serious problems for client/server communications across
different Web-based applications and services. To resolve this problem, the
Open GIS Consortium (OGC) initiated two types of communication
protocol standards for Web-based GIS. The first is the OpenGIS Web
Map Server (WMS) Implementation Interface Specifications, which
provides guidelines for current image-based Internet map servers with
the specifications of HTTP contents and Uniform Resource Locators
(URLs) communication syntax (OGC 2002a). The second standard is the
OpenGIS Web Feature Server (WFS) Implementation Specification. The
Web Feature Service allows a client to retrieve geospatial data encoded in
Geography Markup Language (GML) (vector-based) from multiple Web
Feature Services (OGC 2002b). The WFS adopted XML-based commu-
nication interfaces and GML for describing vector-based features (points,
lines, and polygons). Although the development of WMS and WFS may
solve the potential problem in vendor-based communication protocols,
one issue remains unresolved: how image-based WMS and vector-based
WFS can be integrated. The differences between WMS and WFS reflect
the fundamental challenge of information display, which is the incompat-
ibility between a vector-based map and a raster-based image.
One possible development of client/server communication protocols is to
adopt the Web services frameworks. Currently, Web services comprise the
most exciting developments within Web-based GIS and on-line remote
sensing sources. Web services are formed by the integration of several key
protocols and standards: XML, WSDL (Web Services Definition Language),
SOAP (Simple Object Access Protocol), and UDDI (Universal Description,
Discovery, and Integration). The power of Web services is their combination
of these elements under a single user-friendly operating environment using a
Web-based user interface (Tsou and Buttenfield 2002).
The next section illustrates a interim solution for the integration of both
Web-based GIS display and remotely sensed data. The Internet map
server implementation goal was to adopt a raster-based Internet map
164 Ming-Hsiang Tsou

server (ArcIMS Image server) and JavaScript-based dynamic HTML

documents (DHTML) to display both GIS layers and remote sensing
images. Although this approach is not an optimal solution, it did fulfill
the function of information display and query services from the users’
perspective.

4.2 The implementation of internet map servers

The prototype Web site demonstrated in this project used ESRI’s ArcIMS to
provide interactive Web mapping services. ArcIMS supports OpenGIS WMS
Implementation Interface Specifications 1.0 (ESRI, 2001). The Internet Map
Server provides comprehensive on-line mapping capabilities, including zoom-
in, zoom-out, pan, spatial query, buffering, and measuring. Two types of Web
mapping services (HTML viewer and Java Viewer) were created for different
clusters of GIS tasks. This project also utilizes JavaScript functions to create
a dynamic map display function where users can move the mouse cursor to
overlay different GIS layers on a remotely sensed image (Fig. 3).
One of the unique features in the prototype is the integration of the GPS
data, remotely sensed imagery, and GIS layers using a single Internet Map
Server. This project gathered several GPS data layers, such as park trails and

Fig. 3. The JavaScript-based dynamic Web mapping

Integrating Web-based GIS and image processing tools 165

land cover change sites generated by MTRP park rangers with their hand-
held GPS devices. The GPS data sets were transformed into GIS data format
(shapefiles). These data sets were overlaid with georeferenced ADAR images
(Fig. 4). Park rangers and habitat managers used these Web-based mapping
services to perform various tasks, such as monitoring the change of
landform, querying the locations of sensitive plants, and evaluating the
accessibility of park facilities.
By implementing dynamic Web mapping services, natural resource
managers and park rangers can access geographic information and remotely
sensed data using a typical desktop computer without installing expensive
GIS and remote sensing software packages. Moreover, the Web-based
mapping services become a wonderful education tool for the public learn
about the importance of environmental conservation and habitat manage-
ment. The next section focuses on the third level of integration: spatial
analysis services and their related technologies.

5 Level three: spatial analysis services

5.1 The development of on-line spatial analytical tools

The third level of integration between Web-based GIS and on-line remote
sensing facilities is spatial analysis services, which is the most advanced part
of the integration process. The scope of on-line spatial analysis services
includes Web-based GIS functions (address matching, network analysis,
reselection, etc.) and advanced remotely sensed image analysis (change

Fig. 4. The integration of GPS data, remotely sensed imagery, and GIS layers
166 Ming-Hsiang Tsou

detection, image classification, multiple-bands display, etc.). Different from

the information display and query services, spatial analysis services provides
users the ability to create new information or data by using Web-based
spatial analysis tools or GIS components. There are two major challenges in
this level: ubiquitous access and client/server computation balance.
The first challenge is to provide ubiquitous access to spatial analysis
functions across heterogeneous operating systems and Web browser
environments. Traditional GIS programs and analytical functions are
machine-dependable and vendor-based. To create Web-based GIS modules
or analytical functions, software developers must choose a universal,
interoperable programming environment for implementation. Currently,
several programming languages and techniques, such as Java, ActiveX
controls, and C# (C-sharp), support platform-independent applications
across the Internet (Orfali and Harkey 1997). There is great potential in
utilizing these techniques to create comprehensive Web-based GIS and image
analysis tools. However, some technical issues remain to be resolved for the
full integration of Web-based GIS and remote sensing tools. For example,
the required Application Programming Interfaces (API) and shared pro-
gramming libraries for Web-based GIS might be quite different from remote
sensing applications. Web-based GIS typically focus on the vector-based
computations and the database linkage between spatial features and
attribute records. Remote sensing applications generally require advanced
2D graphics and image functions. To communicate between Web-based GIS
modules and remote sensing applications, standardized API or object
component brokers need to be defined for both software components. Many
software companies and GIS vendors are focusing on this issue, hoping to
come up with a better solution in the near future.
The second challenge is to create a balanced computation load between
client machines and server machines. Since most GIS analysis and remote
sensing image processes require a great amount of computing power, it is
highly desirable to assign the major computation tasks to the most powerful
of the available machines. There are two possible approaches, a server-side
solution or a client-side solution. The server-side solution uses the server to
perform the major computation tasks. The role of clients (as thin clients) is
that of a terminal for sending out users’ requests and receiving final results.
Currently, Web-based Common Gateway Interface (CGI), Java Servlets,
Active Server Page (ASP), and Web Services adopted this type of server-side
solution. The client-side solution puts the major computation tasks on the
client side machine (as thick-clients). The GIS functions or remote sensing
applications are dynamically downloaded into the client side Web browser or
desktop. The role of the server becomes that of a software archive center for
the end users. All major computation and spatial analysis functions happen
on the client side machine. Java applets and ActiveX controls are two
popular types of client-side solutions.
To integrate Web-based GIS components and remote sensing applica-
tions, software developers must consider the nature of their functions and
computation needs. One possible solution is to create a dynamic, LEGO-
like environment for spatial analysis services. The LEGO metaphor refers
to the well-known children’s toy blocks that can be interlocked and
stacked. The LEGO architecture may persist only briefly, for the
Integrating Web-based GIS and image processing tools 167

completion of a single GIS or remote sensing task. Then the LEGO

modules disperse, to be rearranged and restacked in a different config-
uration for a different task (Tsou and Buttenfield 2002). The following
section introduces an actual implementation of Java-based, on-line image
analysis tools for the integration of Web-based GIS functions and remote
sensing applications.

5.2 Implementation of Java-based tools

This project adopted the Java programming for the implementation of on-
line GIS and remote sensing analytical functions. One primary consideration
in choosing a development platform is that the language should provide
database connectivity and comprehensive image processing and display
functions. Java is a pure object-oriented language, designed to enable the
development of secure, high performance, and highly robust applications on
multiple platforms in heterogeneous, distributed networks (Gosling and
McGilton 1996).
Current Java System Development Toolkits (JDK) provides a series of
well-defined APIs for image processing and display, such as Java 2D API
and Java Advanced Imaging (JAI). The Java 2D API is a set of classes for
advanced 2D graphics and imaging, encompassing line art, text, and images
in a single comprehensive model. The Java Advanced Imaging APIs are used
for manipulating and displaying images. They range in complexity from
simple operations, such as contrast enhancement, cropping, and scaling, to
more complex operations such as advanced geometric warping and
frequency domain processing. These APIs are used in a variety of
applications, including geospatial data processing and medical imaging.
This project utilized two APIs (2D and JAI) to customize Web-based user
interface and analytical tools.
There are two procedures for the implementation of Java applets, a
compile-time environment (server-side) and a run-time environment (client-
side). The compile-time environment can be constructed by using the Java
Development Kit (JDK), which include a Java compiler (Javac.exe), a Java
interpreter (Java.exe), a Java debugger (jdb.exe), and several standardized
Java libraries (Krammer 1996). Programmers can use the Java compiler to
generate a Java class from a text-based Java source code to a Java byte-codes
format and put the class on the server-side machine. Then, the Java class is
ready for on demand download by users from client machines.
This project adopts Java applets to provide on-line analytical functions for
remotely sensed image analysis and change detection. The principle reason
for selecting Java applets is that Java applets are specifically designed for the
distributed network environment, such as the Internet and Intranet. They are
capable of providing advanced image processing and comparison functions
suitable for habitat monitoring. One of the unique capabilities of the graphic
user interface (GUI) toolbox design is that it enables users to open multiple
Java windows at the same time (Fig. 5). The ability to open multiple
windows simultaneously gives analysts much greater flexibility in conducting
image comparison and spatial analyses.
There are eight independent Java applets available in the toolbox, which
were originally developed under this project. Examples of representative
168 Ming-Hsiang Tsou

Fig. 5. Multiple Java applets opened from the Image Analysis Toolbox

Java applet source code are listed below. (A detailed description of these
Java source codes can be accessed from the project’s Web site http://
map.sdsu.edu/arc ).
class CompPanel extends JPanel f
Image theImage;
Image theImage2;
public CompPanelðImage theImage; Image theImage2Þ f
this:theImage ¼ theImage;
this:theImage2 ¼ theImage2;
g
public float changeRuleðfloat aÞ f
alpha ¼ a;
swipeWidth ¼ ðintÞða  getSizeðÞ:widthÞ;
System:out:printlnð‘‘a ¼ ’’ þ aÞ;
repaintðÞ;
return a;
g
In this Java example, the analysis function is to overlay two remotely
sensed images (theImage and theImage2) and then switch the two images’
order of display for the purpose of detecting changes (usually temporally)
Integrating Web-based GIS and image processing tools 169

between the two images. The display of images is created by the

[CompPanel] object (Comparison Panel) which has one image analysis
function (changeRule).
Eight different Java applets were implemented in the Image Analysis
Toolbox, including Image Overlay-I, Image Overlay-II, Image Swipe, Image
Magnifier, Image Comparison, Image Processing, Image Filtering, and
Image Viewing. The following paragraphs describe four representative Java
applets and their analytical functions for remote sensing and change
detection analyses.

5.2.1 Image Overlay I and II

Image Overlay-I applet can overlay two remote sensing images and allow a
pixel-by-pixel comparative display of the two image files by means of a slider
bar. This applet enables visual comparison of image changes for monitoring
land covers and habitats. For example, users can compare the 1998 ADAR
imagery with the 1999 ADAR imagery to view land cover changes resulting
from a fire in Mission Trail Regional Park (Fig. 6).
The next Java applet is called ‘‘Image Overlay-II,’’ and is an improved
version of Image Overlay-I. The research team improved the function of
Image Overlay based on the suggestions from user feedback and question-
naires during the prototyping. Image Overlay-II integrates with zoom-and-
scale functions and allows analysts to explore multi-temporal images using
an interactive blend/fade control for improved interpretation of change
detection. Users can move the slider back and forth, and swap top and
bottom images. In addition, the zoom-box can be dragged, enlarged or
reduced to display specific image details. This improved Java applet can
perform more flexible image overlay functions by zooming to a small area on
the image (Fig. 7).

5.2.2 Image Processing

Two Java applets were developed for the purpose of image processing
(Fig. 8). The first Java applet provides basic image processing functions

Fig. 6. The Image Overlay–I Java applet. (showing 1998 and 1999 ADAR images)
170 Ming-Hsiang Tsou

Fig. 7. The Image Overlay II Java applet (with flexible zoom-in capability)

Fig. 8. The Image Processing Java applets

including image enhancement, zoom-in, zoom-out, image smoothing, image

sharpening, image embossing, etc. The second Java applet provided the
capability to display multi-band (RGB) or single-band (gray-scale) imagery.
This project demonstrated the feasibility of an image analysis toolbox. The
demonstration toolbox contained multiple tools defined by eight different
Java applets. Field tests of the toolbox demonstrated the potential capability
of Web-based analytical tools to greatly aid resource and habitat managers
through the analysis of for remote sensed images. Java programmers can
easily customize these Java source codes to create new applications or
combine additional functions. Several research projects and GIS/Remote
Sensing vendors have already adopted Java as their principle development
tool. For example, the GeoVISTA Studio developed at Penn State University
(www.geovistastudio.psu.edu) is one successful example of adopting Java
Integrating Web-based GIS and image processing tools 171

and component-based software engineering techniques to provide a visual

programming environment for geoscientific data analysis and visualization
(Takatsuka and Gahegan, 2002).
The previous sections of the paper have introduced an integrated Web-
based GIS architecture that combining three levels of Web-based geographic
information services (GIServices): data warehouses, Web mapping facilities,
and Java-based image analytical tools. The next section describes the
evaluation of the prototype Websites and their potential development for
environmental monitoring and management.

6 Prototype evaluation and potential developments

A prototype system (http://map.sdsu.edu/arc) was developed using data

warehouses, Internet map servers, and Java-based programming tools. The
final task of this project was to test the prototype to evaluate the utility of its
Web mapping functionality and analytical tools. The research team
conducted two expert review sessions with GIS professionals, park rangers
and GIS and remote sensing graduate students. Training tutorials for the
prototype and evaluation questionnaires were formatted as standardized
PDF files and HTML documents and were posted on the project Web site.
Posting project documentation allowed potential Web users access to the
documents and the ability to respond to the questionnaire following their
evaluation of the prototype system.
In general, the feedback from the evaluation team was very positive. All
participants felt that the Web-based tools were useful for many of their day
to day analysis tasks and clearly have a great potential for improving habitat
management. The following key points were obtained from user feedback.
• The Web-based interfaces are easy to use (comparing to other GIS/
Remote Sensing software).
• The data warehouse needs to clarify some metadata terminology.
• Internet map servers would benefit from more detailed temporal or
seasonal information about remotely sensed images and GIS layers.
• Web-based tools have great potential as an information resource for
Mission Trail Regional Park visitors and the public.
• Viewing and querying GIS data layers was difficult while remotely sensed
image were in the background (via ArcIMS HTML viewers).
Based on the user feedback and the discussions between evaluators and the
research team, two potential enhancements were identified for application to
environmental conservation and habitat monitoring.
The first application enhancement addresses long term change detection
and habitat monitoring needs. Web-based display of multiple-year remotely
sensed imagery can clearly indicate long term temporal changes of land cover
and vegetation. Combined with the GIS layers and spatial analysis functions,
researchers can focus on long term changes in specific habitat areas for future
investigation. From a park ranger’s perspective, Web-based remote sensing
tools can provide important visual clues as to where and when changes are
taking place (change detection). On-line GIS analytical functions can then be
used to determine causes (why) and assist in generating possible solutions for
environmental protection and management tasks (how).
172 Ming-Hsiang Tsou

The second enhancement addresses the desire to combine Web-based GIS

and remote sensing tools with Global Positioning System (GPS) and mobile
devices (Pocket PC or PDA). Evaluator comments repeatedly expressed the
belief that resource managers working in field environments would benefit
from the use of mobile GIS applications (such as ArcPad or MapXtend) to
access remotely sensed imagery and Web-based GIS via Wi-Fi wireless
communication network. Natural habitat preservation managers and scien-
tists can undertake habitat monitoring and change detection tasks in real
time and submit their field observations and reports back to their
organization’s Web servers in near-real time. By integrating GPS, wireless
communication, Web-based GIS and remote sensing tools, park rangers and
other resource mangers can optimize their field-based management tasks
significantly improving field operations efficiency. One such mobile GIS
prototype application has been initiated recently at San Diego State
University (http://map.sdsu.edu/mobilegis).

7 Conclusions

This research and the prototype Web site have demonstrated the potential
and opportunities afforded by the integration of Web-based GIS and on-line
remote sensing facilities for environmental monitoring and management. The
integration of Web-based GIS and remote sensing tools offers significant
benefits over traditional field survey methods associated with resource and
habitat management duties. Traditional approaches for accessing very large
volumes of remotely sensed imagery are time-consuming and problematic.
Only experienced users with sufficient software training can access geospatial
data and imagery through use of complicated GIS and Remote Sensing
software. Before the advent of Web-based GIS, expensive GIS and Remote
Sensing software packages, complicated software installation, and the lack of
software training prevented regional park rangers and local natural resource
managers to utilize the power of GIS and Remote Sensing. By adopting
Web-based GIS and on-line remote sensing facilities, natural resource
managers or regional park rangers can access valuable geospatial informa-
tion and images for their daily tasks without the challenge and cost of
upgrading GIS and image processing software on their local computers.
In summary, Web-based GIS and on-line imagery analysis tools provide a
flexible way to access, extract and create spatial information essential for
environmental monitoring tasks. The need for a capability to monitor
natural preserves using simple change detection methods is worldwide and is
fundamental to cost-effective management practices. This paper has intro-
duced an integrated Web-based GIS architecture by combining three levels
of GIServices: data archive, information display, and spatial analysis. An on-
line data warehouse was developed for archiving, accessing and downloading
both GIS databases and remotely sensed imagery. Multiple interactive map
servers and image servers were established to provide Web-based mapping
functions for the display of land use, vegetation, soil, trails, roads, and
remotely sensed images. Java-based on-line analytical tools (Java applets)
provided advanced image comparison and analysis functions for land cover
change detection. This paper also discussed the major challenges to
integrating Web-based GIS and on-line image analysis tools, including
Integrating Web-based GIS and image processing tools 173

metadata standards, data transmission formats, communication protocols,

client/server balance, and ubiquitous access. Hopefully, these problems will
be solved and the challenges answered by continued collaborations between
the GIS and Remote Sensing communities in the foreseeable future.

References

Buehler K, McKee L (eds) (1998) The Open GIStm Guide: Introduction to Interoperable
Geoprocessing and the OpenGIS Specification (Third Edition). Open GIS Consortium, Inc.,
Wayland, Massachusetts
Buttenfield BP, Goodchild MF (1996) The Alexandria Digital Library Project: distributed
library services for spatially referenced data. In: Proceedings of GIS/LIS’96, Denver,
Colorado, November 19–21, 1996, ACSM, pp 76–84
Ehlers M, Edwards G, Bedard Y (1989) Integration of Remote Sensing with Geographic
Information Systems: A Necessary Evolution. Photogrammetric Engineering & Remote Sensing
55: 1619–1627
Elkington M, Meyer R, McConaughy G (1994) Defining the Architectural Development of
EOSDIS to Facilitate Extension to a Wider Data Information System. NASA Technical
Paper, April 1994. EOSDIS Core System Project. 194-00131. NASA, Landover, Maryland
Environmental Systems Research Institute (ESRI) (2001) Using ArcIMS. (User’s Manual).
ESRI, Redland, California
Estes JE, Star JL (1993) Remote Sensing and GIS Integration: Towards a Prioritized Research
Agenda. NCGIA Technical Report 93-4, May 1993, NCGIA, Santa Barbara, California
Federal Geographic Data Committee (FGDC) (1998) Content Standards for Digital Geospatial
Metadata (revised June 1998). FGDC-STD-001-1998. FGDC, Washington, D.C.
Federal Geographic Data Committee (FGDC) (2002) Content Standards for Digital Geospatial
Metadata: Extensions for Remote Sensing Metadata. FGDC-STD-012-2002, FGDC/USGS,
Reston, Virginia
Fonseca F, Egenhofer M, Agouris P, Câmara G (2002) Using Ontologies for Integrated
Geographic Information Systems. Transactions in GIS (6) 3: 231–257
Frew J, Freitas N, Hill L, Lovette K, Nideffer R, Zheng Q (1998) The Alexandria Digital Library
System Architecture. In: Strobel J, Best C (eds) Proceedings of the Earth Observation Geo-
Spatial Web and Internet Workshop ’98 (Salzburger Geographische Materialien, vol. 27).
Salsburg, Instituts für Geographie der Universität Salzburg. URL: http://www.sbg.ac.at/geo/
eogeo/authors/frew/frew.htm (last access date: 9-29-2003)
Gosling J, McGilton H (1996) The Java Language Environment, A White Paper. Sun
Microsystems, Santa Clara, California. URL: http://www.Java.sun.com/docs/white/langenv/
(last access date: 9-29-2003).
Gupta A, Marciano R, Zaslavsky I, Baru C (1999) Integrating GIS and Imagery through
XML-Based Information Mediation. In: Agouris P, Stefanidis A (eds) Integrated Spatial
Databases - Digital Images and GIS, Lecture Notes in Computer Science 1737 Springer, Berlin,
pp 211–234
ISO/TC 211/WG 3 (1998) Geographic Information – Part 15: Metadata. ISO/TC 211-N538, ISO/
CD 15046-15, International Standard Organization, Geneva, Switzerland
Kemp Z (1999) A framework for integrating gis and images. In Agouris P, Stefanidis A (eds)
Integrated Spatial Databases - Digital Images and GIS, Lecture Notes in Computer Science
1737 Springer, Berlin, pp 153–167
Krammer D (1996) The Java Platform, A White Paper, Sun Microsystems, Santa Clara,
California. URL: http://java.sun.com/docs/white/platform/javaplatformTOC.doc.html (last
access date: 9-29-2003)
Li B, Zhang L (1997) A Model of Component-oriented GIS. In: Proceedings of GIS/LIS’97,
October 28-30, 1997, Cincinnati, Ohio, ACSM, pp 523–528
NASA (2002) SEEDS Standards and Interface Processes Draft Survey Report. (Draft Version
1.9.2) June 14, 2002, NASA, Landover, Maryland
174 Ming-Hsiang Tsou

Open GIS Consortium, Inc. (OGC) (2002a). OpenGIS Web Map Server Interface Implementation
Specification (Version 1.1.1). Open GIS Consortium, Inc., Wayland, Massachusetts, URL:
http://www.opengis.org/techno/implementation.htm (last access date: 9-29-2003)
Open GIS Consortium, Inc. (OGC) (2002b). OpenGIS Web Feature Server Interface Implemen-
tation Specification (Version 1.0.0). Open GIS Consortium, Inc., Wayland, Massachusetts,
URL: http://www.opengis.org/techno/implementation.htm (last access date: 9-29-2003)
Open GIS Consortium, Inc. (OGC) (2003). OpenGIS Geography Markup Language (GML)
Implementation Specification (Version 3.0). Open GIS Consortium, Inc., Wayland, Massa-
chusetts, URL: http://www.opengis.org/techno/implementation.htm (last access date: 9-29-
2003)
Orfali R, Harkey D (1997) Client/Server Programming with Java and CORBA. John Wiley &
Sons, Inc, New York
Ostensen O (1995) Mapping the Future of Geomatics. ISO Bulletin, December 1995: 13–15
Putz S (1994) Interactive Information Services Using World Wide Web Hypertext. In
Proceedings of the First International Conference on the World-Wide Web, May 25-27, 1994,
Geneva, Switzerland
Takatsuka M, Gahegan M (2002) GeoVISTA Studio: A Codeless Visual Programming
Environment for Geoscientific Data Analysis and Visualization. Computers & Geosciences
28(10): 1131–1144
Tsou MH (2002) An Operational Metadata Framework for Searching, Indexing, and Retrieving
Distributed Geographic Information Services on the Internet. In: Egenhofer M, Mark D (eds)
Geographic Information Science-Second International Conference GIScience 2002, Lecture
Notes in Computer Science 2478, Springer, Berlin, pp 312–333
Tsou MH, Buttenfield BP (2002) A Dynamic Architecture for Distributing Geographic
Information Services. Transactions in GIS 6(4): 355–381
Vckovski A (1998) Interoperable and Distributed Processing in GIS. Taylor & Francis, London
Zaslavsky I (2000) A New Technology for Interactive Online Mapping with Vector Markup and
XML. Cartographic Perspectives 37: 65–77